Application of Deep-learning Methods in the Classification of Chinese Buddhist Canonical Catalogs

Abstract

This research focuses on the classification of Chinese Buddhist scripture catalogs, employing advanced deep learning methods to develop an automatic classification mechanism. Chinese Buddhist scripture catalogs serve as vital tools for organizing and retrieving Buddhist literature, facilitating research in the field. However, with the continuous addition of new texts and the need to adapt to modern academic research requirements, traditional manual classification methods have become time-consuming and less effective. In this study, we aim to address this challenge by harnessing the power of deep learning techniques. Our research not only involves the integration of new literature into existing catalog structures but also explores the reasons behind misclassifications in these catalogs. Additionally, we examine the inherent connections between different categories of scriptures to provide a comprehensive understanding of the catalog structure. Our contributions in this research encompass: 1. Automatic Classification: We pioneer the use of deep learning methods for the classification of Buddhist scripture catalogs, allowing for faster and more accurate categorization of texts. This automated approach significantly enhances the efficiency of catalog management. 2. Error Analysis: We delve into the reasons behind misclassifications in the catalogs, shedding light on common pitfalls and misconceptions in traditional classification methods. 3. Interconnections: We uncover the original interrelationships between different categories of scriptures, offering valuable insights for adjusting and optimizing the catalog structure to align with contemporary research needs. 4. Practical Application: The research findings are adopted by the CBETA research group as a reference tool for the cataloging of new literature in future editions of the Tripitaka, ensuring that the catalog remains up-to-date and relevant. This research is a significant step towards modernizing the management of Chinese Buddhist scripture catalogs, making them more efficient, accurate, and adaptable to the evolving landscape of Buddhist literature. (Article content in Chinese with English extended abstract)

Keywords

• cbeta
• tripitaka
• scripture catalog
• deep-learning methods
• text classification